A model-based approach of data analysis and prediction in cardiovascular disease

During recent years, the number of machine learning and data mining algorithms applied into healthcare systems for disease’s diagnosis and prediction has increased significantly. Their effective application can lead to significant costs savings related with treatment management, hospital admissions and drugs. This research targets cardiovascular disease, which is a serious chronic disease that affects the heart’s performance and the individual’s quality of life. Cardiovascular diseases can be predicted based on the patient’s health attributes and conditions. In the pursuit of improving the diagnosis process of cardiovascular disease, multiple data mining models have been proposed on the literature. The developed models in this study are based on the Cardiovascular Disease dataset published by Ulianova. This study suggests a model-based framework that utilizes multiple data mining techniques to be used as decision support tools in the diagnosis of cardiovascular diseases. The proposed model, after the training and testing of multiple classification methods, utilizes a random forest as a prime classifier that can help researchers and physicians improve the prediction process of cardiovascular diseases, and thus, to get more accurate results and better healthcare outcomes for patients. The validation of the developed models shows accuracy, sensitivity, recall and F-measure results.